Chapter 8

Roman Jakobson

ON LINGUISTIC ASPECTS OF

TRANSLATION

AC C O R D I N G T O B E RT R A N D R U S S E L L , “no one can understandthe word ‘cheese’ unless he has a nonlinguistic acquaintance with cheese.”1 If,

however, we follow Russell’s fundamental precept and place our “emphasis uponthe linguistic aspects of traditional philosophical problems,” then we are obliged tostate that no one can understand the word “cheese” unless he has an acquaintancewith the meaning assigned to this word in the lexical code of English. Anyrepresentative of a cheese-less culinary culture will understand the English word“cheese” if he is aware that in this language it means “food made of pressed curds”and if he has at least a linguistic acquaintance with “curds.” We never consumedambrosia or nectar and have only a linguistic acquaintance with the words“ambrosia,” “nectar,” and “gods”—the name of their mythical users; nonetheless,we understand these words and know in what contexts each of them may be used.

The meaning of the words “cheese,” “apple,” “nectar,” “acquaintance,” “but,”“mere,” and of any word or phrase whatsoever is definitely a linguistic—or to bemore precise and less narrow—a semiotic fact. Against those who assign meaning(signatum) not to the sign, but to the thing itself, the simplest and truest argumentwould be that nobody has ever smelled or tasted the meaning of “cheese” or of“apple.” There is no signatum without signum. The meaning of the word “cheese”cannot be inferred from a nonlinguistic acquaintance with cheddar or withcamembert without the assistance of the verbal code. An array of linguistic signs isneeded to introduce an unfamiliar word. Mere pointing will not teach us whether“cheese” is the name of the given specimen, or of any box of camembert, or ofcamembert in general or of any cheese, any milk product, any food, anyrefreshment, or perhaps any box irrespective of contents. Finally, does a word

1959

114 ROMAN JAKOBSON

simply name the thing in question, or does it imply a meaning such as offering,sale, prohibition, or malediction? (Pointing actually may mean malediction; insome cultures, particularly in Africa, it is an ominous gesture.)

For us, both as linguists and as ordinary word-users, the meaning of anylinguistic sign is its translation into some further, alternative sign, especially asign “in which it is more fully developed,” as Peirce, the deepest inquirer into theessence of signs, insistently stated.2 The term “bachelor” may be converted into amore explicit designation, “unmarried man,” whenever higher explicitness isrequired. We distinguish three ways of interpreting a verbal sign: it may betranslated into other signs of the same language, into another language, or intoanother, nonverbal system of symbols. These three kinds of translation are to bedifferently labeled: 1 Intralingual translation or rewording is an interpretation of verbal signs by

means of other signs of the same language.2 Interlingual translation or translation proper is an interpretation of verbal

signs by means of some other language.3 Intersemiotic translation or transmutation is an interpretation of verbal signs

by means of signs of nonverbal sign systems. The intralingual translation of a word uses either another, more or lesssynonymous, word or resorts to a circumlocution. Yet synonymy, as a rule, is notcomplete equivalence: for example, “every celibate is a bachelor, but not everybachelor is a celibate.” A word or an idiomatic phrase-word, briefly a code-unitof the highest level, may be fully interpreted only by means of an equivalentcombination of code-units, i.e., a message referring to this code-unit: “everybachelor is an unmarried man, and every unmarried man is a bachelor,” or“every celibate is bound not to marry, and everyone who is bound not to marryis a celibate.”

Likewise, on the level of interlingual translation, there is ordinarily no fullequivalence between code-units, while messages may serve as adequateinterpretations of alien code-units or messages. The English word “cheese” cannotbe completely identified with its standard Russian heteronym “ ,” because cottagecheese is a cheese but not a . Russians say: “bringcheese and [sic] cottage cheese.” In standard Russian, the food made of pressedcurds is called only if ferment is used.

Most frequently, however, translation from one language into another substitutesmessages in one language not for separate code-units but for entire messages insome other language. Such a translation is a reported speech; the translator recodesand transmits a message received from another source. Thus translation involvestwo equivalent messages in two different codes.

Equivalence in difference is the cardinal problem of language and the pivotalconcern of linguistics. Like any receiver of verbal messages, the linguist acts astheir interpreter. No linguistic specimen may be interpreted by the science oflanguage without a translation of its signs into other signs of the same system orinto signs of another system. Any comparison of two languages implies anexamination of their mutual translatability; widespread practice of interlingual

ON LINGUISTIC ASPECTS OF TRANSLATION 115

communication, particularly translating activities, must be kept under constantscrutiny by linguistic science. It is difficult to overestimate the urgent need for andthe theoretical and practical significance of differential bilingual dictionaries withcareful comparative definition of all the corresponding units in their intention andextension. Likewise differential bilingual grammars should define what unifies andwhat differentiates the two languages in their selection and delimitation ofgrammatical concepts.

Both the practice and the theory of translation abound with intricacies, andfrom time to time attempts are made to sever the Gordian knot by proclaimingthe dogma of untranslatability. “Mr. Everyman, the natural logician,” vividlyimagined by B.L.Whorf, is supposed to have arrived at the following bit ofreasoning: “Facts are unlike to speakers whose language background providesfor unlike formulation of them.”3 In the first years of the Russian revolution therewere fanatic visionaries who argued in Soviet periodicals for a radical revisionof traditional language and particularly for the weeding out of such misleadingexpressions as “sunrise” or “sunset.” Yet we still use this Ptolemaic imagerywithout implying a rejection of Copernican doctrine, and we can easily transformour customary talk about the rising and setting sun into a picture of the earth’srotation simply because any sign is translatable into a sign in which it appears tous more fully developed and precise.

A faculty of speaking a given language implies a faculty of talking about thislanguage. Such a “metalinguistic” operation permits revision and redefinition ofthe vocabulary used. The complementarity of both levels—object-language andmetalanguage—was brought out by Niels Bohr: all well-defined experimentalevidence must be expressed in ordinary language, “in which the practical use ofevery word stands in complementary relation to attempts of its strict definition.”4

All cognitive experience and its classification is conveyable in any existinglanguage. Whenever there is deficiency, terminology may be qualified and amplifiedby loan-words or loan-translations, neologisms or semantic shifts, and finally, bycircumlocutions. Thus in the newborn literary language of the Northeast SiberianChukchees, “screw” is rendered as “rotating nail,” “steel” as “hard iron,” “tin” as“thin iron,” “chalk” as “writing soap,” “watch” as “hammering heart.” Evenseemingly contradictory circumlocutions, like “electrical horse-car”( ), the first Russian name of the horseless street car, or “flyingsteamship” (jena paragot), the Koryak term for the airplane, simply designate theelectrical analogue of the horse-car and the flying analogue of the steamer and donot impede communication, just as there is no semantic “noise” and disturbance inthe double oxymoron—“cold beef-and-pork hot dog.”

No lack of grammatical device in the language translated into makes impossiblea literal translation of the entire conceptual information contained in the original.The traditional conjunctions “and,” “or” are now supplemented by a newconnective—“and/or”—which was discussed a few years ago in the witty bookFederal Prose—How to Write in and/or for Washington.5 Of these threeconjunctions, only the latter occurs in one of the Samoyed languages.6 Despitethese differences in the inventory of conjunctions, all three varieties of messagesobserved in “federal prose” may be distinctly translated both into traditional Englishand into this Samoyed language. Federal prose: 1) John and Peter, 2) John or Peter,

116 ROMAN JAKOBSON

3) John and/or Peter will come. Traditional English: 3) John and Peter or one ofthem will come. Samoyed: John and/or Peter both will come, 2) John and/or Peter,one of them will come.

If some grammatical category is absent in a given language, its meaning may betranslated into this language by lexical means. Dual forms like Old Russian ?paraare translated with the help of the numeral: “two brothers.” It is more difficult toremain faithful to the original when we translate into a language provided with acertain grammatical category from a language devoid of such a category. Whentranslating the English sentence “She has brothers” into a language whichdiscriminates dual and plural, we are compelled either to make our own choicebetween two statements “She has two brothers”—“She has more than two” or toleave the decision to the listener and say: “She has either two or more than twobrothers.” Again in translating from a language without grammatical number intoEnglish one is obliged to select one of the two possibilities—“brother” or “brothers”or to confront the receiver of this message with a two-choice situation: “She haseither one or more than one brother.”

As Boas neatly observed, the grammatical pattern of a language (as opposedto its lexical stock) determines those aspects of each experience that must beexpressed in the given language: “We have to choose between these aspects, andone or the other must be chosen.”7 In order to translate accurately the Englishsentence “I hired a worker,” a Russian needs supplementary information, whetherthis action was completed or not and whether the worker was a man or a woman,because he must make his choice between a verb of completive or noncompletiveaspect— or —and between a masculine and feminine noun—

or . If I ask the utterer of the English sentence whether theworker was male or female, my question may be judged irrelevant or indiscreet,whereas in the Russian version of this sentence an answer to this question isobligatory. On the other hand, whatever the choice of Russian grammatical formsto translate the quoted English message, the translation will give no answer tothe question of whether I “hired” or “have hired” the worker, or whether he/shewas an indefinite or definite worker (“a” or “the”). Because the informationrequired by the English and Russian grammatical pattern is unlike, we face quitedifferent sets of two-choice situations; therefore a chain of translations of one andthe same isolated sentence from English into Russian and vice versa could entirelydeprive such a message of its initial content. The Geneva linguist S.Karcevskiused to compare such a gradual loss with a circular series of unfavorable currencytransactions. But evidently the richer the context of a message, the smaller theloss of information.

Languages differ essentially in what they must convey and not in what they mayconvey. Each verb of a given language imperatively raises a set of specific yes-or-no questions, as for instance: is the narrated event conceived with or withoutreference to its completion? Is the narrated event presented as prior to the speechevent or not? Naturally the attention of native speakers and listeners will beconstantly focused on such items as are compulsory in their verbal code.

In its cognitive function, language is minimally dependent on the grammaticalpattern because the definition of our experience stands in complementary relationto metalinguistic operations—the cognitive level of language not only admits but

ON LINGUISTIC ASPECTS OF TRANSLATION 117

directly requires receding interpretation, i.e., translation. Any assumption ofineffable or untranslatable cognitive data would be a contradiction in terms. But injest, in dreams, in magic, briefly, in what one would call everyday verbal mythologyand in poetry above all, the grammatical categories carry a high semantic import.In these conditions, the question of translation becomes much more entangled andcontroversial.

Even such a category as grammatical gender, often cited as merely formal,plays a great role in the mythological attitudes of a speech community. In Russianthe feminine cannot designate a male person, nor the masculine specify a female.Ways of personifying or metaphorically interpreting inanimate nouns are promptedby their gender. A test in the Moscow Psychological Institute (1915) showed thatRussians, prone to personify the weekdays, consistently represented Monday,Tuesday, and Thursday as males and Wednesday, Friday, and Saturday as females,without realizing that this distribution was due to the masculine gender of thefirst three names ( ) as against the feminine genderof the others ( ). The fact that the word for Friday ismasculine in some Slavic languages and feminine in others is reflected in the folktraditions of the corresponding peoples, which differ in their Friday ritual. Thewidespread Russian superstition that a fallen knife presages a male guest and afallen fork a female one is determined by the masculine gender of “knife”and the feminine of “fork” in Russian. In Slavic and other languages where“day” is masculine and “night” feminine, day is represented by poets as thelover of night. The Russian painter Repin was baffled as to why Sin had beendepicted as a woman by German artists: he did not realize that “sin” is femininein German (die Sünde), but masculine in Russian (Γpex). Likewise a Russianchild, while reading a translation of German tales, was astounded to find thatDeath, obviously a woman (Russian , fem.), was pictured as an old man(German der Tod, masc.). My Sister Life, the title of a book of poems by BorisPasternak, is quite natural in Russian, where “life” is feminine , but wasenough to reduce to despair the Czech poet Josef Hora in his attempt to translatethese poems, since in Czech this noun is masculine z∨∨∨∨∨ivot.

What was the initial question which arose in Slavic literature at its verybeginning? Curiously enough, the translator’s difficulty in preserving the symbolismof genders, and the cognitive irrelevance of this difficulty, appears to be the maintopic of the earliest Slavic original work, the preface to the first translation of theEvangeliarium, made in the early 860’s by the founder of Slavic letters and liturgy,Constantine the Philosopher, and recently restored and interpreted by A.Vaillant.8

“Greek, when translated into another language, cannot always be reproducedidentically, and that happens to each language being translated,” the Slavic apostlestates. “Masculine nouns as ‘river’ and ‘star’ in Greek, are femininein another language as and in Slavic.” According to Vaillant’scommentary, this divergence effaces the symbolic identification of the rivers withdemons and of the stars with angels in the Slavic translation of two of Matthew’sverses (7:25 and 2:9). But to this poetic obstacle, Saint Constantine resolutelyopposes the precept of Dionysius the Areopagite, who called for chief attention tothe cognitive values ( ) and not to the words themselves.

In poetry, verbal equations become a constructive principle of the text. Syntactic

118 ROMAN JAKOBSON

and morphological categories, roots, and affixes, phonemes and their components(distinctive features)—in short, any constituents of the verbal code—are confronted,juxtaposed, brought into contiguous relation according to the principle of similarityand contrast and carry their own autonomous signification. Phonemic similarity issensed as semantic relationship. The pun, or to use a more erudite, and perhapsmore precise term—paronomasia, reigns over poetic art, and whether its rule isabsolute or limited, poetry by definition is untranslatable. Only creativetransposition is possible: either intralingual transposition—from one poetic shapeinto another, or interlingual transposition—from one language into another, orfinally intersemiotic transposition—from one system of signs into another, e.g.,from verbal art into music, dance, cinema, or painting.

If we were to translate into English the traditional formula Traduttore, traditoreas “the translator is a betrayer,” we would deprive the Italian rhyming epigram ofall its paronomastic value. Hence a cognitive attitude would compel us to changethis aphorism into a more explicit statement and to answer the questions: translatorof what messages? betrayer of what values?

Notes

1 Bertrand Russell, “Logical Positivism,” Revue Internationale de Philosophie,IV (1950), 18; cf. p. 3.

2 Cf. John Dewey, “Peirce’s Theory of Linguistic Signs, Thought, and Meaning,”The Journal of Philosophy, XLIII (1946), 91.

3 Benjamin Lee Whorf, Language, Thought, and Reality (Cambridge, Mass.,1956), p. 235.

4 Niels Bohr, “On the Notions of Causality and Complementarity,” Dialectica,I (1948), 317f.

5 James R.Masterson and Wendell Brooks Phillips, Federal Prose (Chapel Hill,N.C., 1948), p. 40f.

6 Cf. Knut Bergsland, “Finsk-ugrisk og almen språkvitenskap,” Norsk Tidsskriftfor Sprogvidenskap, XV (1949), 374f.

7 Franz Boas, “Language,” General Anthropology (Boston, 1938), pp. 132f.8 André Vaillant, “Le Préface de l’Évangeliaire vieux-slave,” Revue des Études

Slaves, XXIV (1948), 5f.

The Translation Studies Reader

Edited by

Lawrence Venuti

Advisory Editor: Mona Baker

London and New York

First published 2000by Routledge11 New Fetter Lane, London EC4P 4EE

Simulataneously published in the USA and Canadaby Routledge29 West 35th Street, New York, NY 10001 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2004. © 2000 This collection and editorial matter © Lawrence Venuti;individual essays © individual contributors

All rights reserved. No part of this book may be reprinted or reproduced or utilised in anyform or by any electronic, mechanical, or other means, now known or hereafter invented,including photocopying and recording, or in any information storage or retrieval system,without permission in writing from the publishers. British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication DataThe Translation studies reader/edited by Lawrence Venuti.

p. cm.Includes bibliographical references and index.1. Translating and interpreting. I. Venuti, Lawrence.

P306.T7436 2000418'.02–dc21 99–36161

CIP ISBN 0-203-44662-3 Master e-book ISBN ISBN 0-203-75486-7 (Adobe eReader Format)ISBN 0-415-18746-X (Hbk)ISBN 0-415-18747-8 (Pbk)